Introduction
The Roman world enjoyed a vibrant clothing culture. Roman dress offered a significant means of expressing identity and the clothed body was a highly symbolic medium. Dress served a functional purpose in protecting and covering the body, but it was also an important area of social compe-tition and display. The wearer manipulated their clothing to signal important associations, promoting, to an extent, how they wished to be perceived.1 This becomes particu-larly apparent from the 3rd century onwards when dressing fashions became increasingly more colourful and ornate. A multiplicity of interactions – from actual experiences to ide-alised interactions – conveyed aspects of dress experience, mediated through different media. It was not just the textiles themselves that negotiated and framed an individual’s rec-ognition of the language of dress. Many versions of dress served to structure and enhance an individual’s engagement with clothing practices. The majority of Roman textiles have not survived, except in cases of extreme environmental con-ditions.2 This material lacuna accurately reflects neither the ubiquity of textile production nor the integral part played by clothing in the diverse and dynamic cultural system that
was such an integral part of identity construction.3 Given this absence of material evidence, any study of ancient dress has to look to other categories of evidence. One such cate-gory is the visual representation of what can be described as ‘image-clothing’. Dress imagery can never, of course, offer the rich details that textile fragments might, but they nevertheless provide some insights into aspects of Roman textile reality, at least visually, and thereby enhance the understanding of the language of clothing fashions.
This paper focuses on one type of visual media: mosa-ics. It examines two 4th-century mosaics from Tunisia: the Dominus Julius Mosaic from Carthage and a villa rustica mosaic from Tabarka. Mosaics offer a far more durable example of how people chose to portray themselves than do fragile textiles from archaeological contexts. Composed of vivid polychrome images, mosaics hint at a colourful textile world. Indeed, the use of increasingly vibrant textiles from the 4th century CE onwards mirrors the growing popularity of colour decor in elite residences.4 But how far does the dress imagery found in mosaics relate to what was worn in real life? Do these mosaics present an ‘accurate’ view of clothing garments or are they merely highly stylised Abstract
This paper examines the iconography in two mosaics: the Dominus Julius Mosaic from Carthage and a villa rustica mosaic from Tabarka, Tunisia. It explores the methodological issues of extrapolating Roman dress practice from mosaics. Depictions of dress in mosaic floors can be understood as a form of ‘represented’ clothing, a form that documents dominant societal ideas and ideals that have been transformed into the visual medium to become
‘image-clothing’. The use of mosaic pavements formed part of broader mechanisms of identity negotiation and expression, particularly in elite circles. In such dress imagery, elite patrons signalled appropriate participation in clothing practice both by employing suitable themes and in the dress styles they were shown wearing. Social differentiation could also be maintained by contrasting how different figures were clothed. ‘Reading’ dress from mosaic imagery requires an appreciation of how visual imagery was intended to be viewed. It was, most signifi-cantly, a version of ‘lived’ dress practice manipulated to reflect contemporary visual conventions.
Amy Place 180
representations? If the latter is the case, then how can such images be used to study Roman dress? Despite the somewhat surprising neglect of mosaic iconography as a resource for the exploration of Roman clothing habits, such evidence can, when studied closely, shed new light on the language of Roman dress.5 Mosaics do this not by offering a pho-tographic snapshot of Roman clothing as it was worn in normal life; instead, elements of Roman sartorial discourse were embedded in imagery and articulated tile by minute tile. It follows that mosaic dress iconography records the painstaking construction of a particular version of dress practice that can be used to decipher both the symbolic and semiotic role of clothing in the later Roman world.
The research begins by outlining some key methodolog-ical issues that arise when using mosaic pavement imagery to interpret contemporary later Roman dress, given that such images may not represent the actual clothing worn in real life. Its purpose is to demonstrate that, by adopting the correct approach, the visual portrayal of dress imagery in mosaic iconography can inform discussions about Roman dress. Two contemporary case studies, the Dominus Julius Mosaic from Carthage, c. 380–400 CE, and a villa rustica mosaic from Tabarka, c. 400 CE, both from Tunisia, North Africa, are used to contextualise the textile iconography that features figurative imagery.6 All too often, modern schol-arship dismisses mosaic pavements as being mere images, the illustrations to a text, and as such divorced from the material significance they once held. Their role in terms of identity construction and elite performance, however, influenced how people at the time interpreted their ico-nography. Conditioned through experience to interpret the iconography in particular ways, the responses of people at the time to the imagery would not have been the same as those of the modern viewer.7 By examining the relationship between dress iconography and clothing discourse it is possible to appreciate both how such imagery contributed to the cultural rhythms of dress in the later Roman world and how dress iconography operated within the bounds of the visual rhetoric of dress.
Visual imagery as a source for Roman dress:
Some methodological issues
Sources for Roman dress essentially fall into three groups:
archaeological, visual and textual. Although each is undoubt-edly informative in its own right, it would be wrong simply to combine evidence from each source without considering each within its own context. Different versions of clothing reveal different interpretations of clothing as worn in their respective circumstances. As a symbolic representational system, clothed bodies reflect prominent ideas and ideals.
This is true for any form of dress, be it metaphorical clothing in literature, visual portrayals or the dressed individual who appears in public. On occasion, the resulting interpretations
may appear to contradict each other as they seek to recognise different aspects of the clothing experience, but this does not necessarily invalidate the accuracy of these sartorial con-clusions. Therefore, investigating the relationship between visual depictions of clothing and the version of dress that the image represents is a crucial step towards interpreting how ancient iconography, and more specifically mosaics, can be utilised in discussions about Roman dress.
How the three categories of evidence for dress, the archaeological, visual and textual, interrelate with each other from a theoretical perspective is a subject tackled by Roland Barthes in Système de la Mode (1967). In his study, he examines contemporary French fashion using evidence from fashion magazines. Barthes’ structural framework offers a useful departure point for dress historians as it provides a methodological foundation for conceptualising different forms of dress evidence and understanding their relationship to one another. His fashion theory can readily be adapted for the study of ancient textiles and his tripartite structure for sources of dress has recently been employed with increasing enthusiasm in the study of Roman dress.8
For Barthes, clothing has three modalities: real clothing, clothing portrayed through images (‘image-clothing’) and clothing that is described in text (‘written clothing’).9 This identification of the three constituent parts of clothing into three ‘versions’ of a garment is perhaps the most influen-tial aspect of Barthes’ formulation of fashion as a system:
his tripartite scheme can also be neatly mapped onto the types of dress evidence used by dress historians.10 The first of these categories refers to the actual textile itself – the
‘technological’ garment – while the other two – the ‘iconic’
and ‘verbal’ garments – can be understood as ‘represented’
clothing. Importantly, though, these two groups of repre-sented clothing do not operate in an identical way; one is a system of images, the other a system of language. In fact, while they refer to the same reality – that is, dress experience – image-clothing and written clothing can inherently convey different ideas and details about the
‘original’ real garment and in so doing overwrite these ideas.
Furthermore, they are not identical to the physical textile itself and can never fully encapsulate the reality of the actual garment; instead, they serve to signify different facets of the original textile and thus offer the modern scholar insights into the rhetoric of contemporary dress.
Knowledge of the structural and technical practicalities of Roman clothing has been enhanced by the developments in Roman textile research. Working from an extensive body of preserved archaeological fragments, this research throws fresh light on the material reality of Roman dress, offering insight into such things as yarn type, weaving technologies, dyeing techniques and decoration.11 These details are important when exploring Roman dress. Yet, as Barthes’ clothing system demonstrates, the technicalities of the garment run along parallel lines to that of the clothing
13. Reading dress and identity in the Roman mosaics of Carthage and Tabarka 181
portrayed in the iconography. This paper focuses on the way clothing has been represented, on the translation of the garment as worn on the body into visual representations of the clothed body. Given how few textiles survive within the archaeological record in large parts of the Roman world, other dress sources inevitably come to the fore. This in not in itself a disadvantage: as Caroline Vout notes, the study of Roman dress is ‘not a study of the clothes themselves but of the images of clothes, not a study of how Roman people looked but of how they perceived or defined themselves as looking’.12 Ursula Rothe draws a similar conclusion in her study of 1st- to 4th-century CE funerary monuments in the Rhine-Moselle region, arguing that elements of selection and standardisation do not preclude these images being at the same time artefacts of how the people viewed themselves.13
According to Barthes’ framework, the signals given out by ‘represented’ garments are governed by their respective structures. How Barthes constructs that interpretative struc-ture is therefore a significant part of his methodology. For written clothing, the transformation of textiles into language through description or a photograph caption limits the num-ber of potential meanings: ‘words determine a single cer-tainty’.14 Barthes’ comment referred to fashion photographs, a genre bound by its own agenda and language. As such, the written clothing which evokes the original technological garment (Barthes calls this the ‘Fashion description’) must describe this real clothing accordingly.15 In the context of ancient textiles, written clothing is more akin to Barthes’
‘literary description’ where references to ‘real’ garments – whether through specific textile terminology, garment names or recognisable associations – are needed to make the garment exist for the reader. In many cases, though, the primary intention behind the written clothing discourse was not necessarily to debate dressing habits per se. The writings of authors like Tertullian or Cyprian of Carthage make the overt language of clothing a proxy for debating wider social anxieties. An example of this is the debate about gender dynamics achieved through Patristic recommendations of female dress practices, particularly those that promote female modesty in both dress and action.16
Effectively, verbal clothing reveals relevant socio-cul-tural associations through reference to specific articles of clothing or modes of dress. Visual media works similarly but offers far more possible meanings for image-clothing than does text, with its finite number of possible meanings. As Barthes writes, ‘we know that in fact an image inevitably involves several levels of perception … every glance cast at an image inevitably implies a decision i.e. the meaning of an image is never certain … The image freezes an endless number of possibilities’.17 Of course, the iconic garment created does not have to exist in a physical sense: it must merely be acknowledged by its audience as a reflection of cultural norms and practice. A pertinent example of this is the proliferation of Roman statues clad in togas as a result
of the ‘public and status-orientated nature of Roman art’.18 As static modes of expression, these cultural artefacts per-formed a particular function and were subject to stylistic conventions and social expectations.19 Public visual imagery of this kind displayed clothing practices deemed suitable for public consumption.20 Although they might offer something worthy of imitation by reinforcing social and gender ideals, it is important to remember that ‘it was never the primary aim of any artistic medium to render detailed representations of either Roman clothing, or of any other textiles’.21 Visual dress was not intended to be copied in everyday life; instead, the ideas embedded within the imagery were meant to be absorbed and emulated by the wider social body.22 This does not diminish the contribution made by visual media to the discussion of Roman textiles; instead, it thereby becomes an artefact that sheds light upon how the Romans visualised clothing, how the viewers responded to such depictions and, most significantly, how visual clothing functioned as a means of communication.23
There are, therefore, multiple versions of Roman dress rhetoric reflected in the different types of sources. These discourses take many different forms and run parallel to one another but at the same time still bear some relationship to each other insofar as they originate from the same composite system, i.e. the social reality of Roman sartorial fashions.
Dress imagery represents clothing practice translated into the visual sphere. That being so, the image-conscious clothing portrayed in mosaics was surely governed by the relevant traditions of that artistic genre which, in turn, conditioned how image-clothing was constructed, displayed and interpreted. Ultimately, mosaic pavements are not exact records of the later Roman textile reality: this would be a far too literal interpretation of their imagery. Instead, such iconography should be appreciated as artefacts of a clothing discourse which reflects visual clothing as experienced in the context of contemporary societal discourse.
Clothing imagery in North African mosaics:
Two case studies
Two well-known North African mosaics, the Dominus Julius Mosaic and the Tabarka villa rustica mosaic panel, both discussed below, illustrate the methodological issues involved in synthesising visual depictions of dress. These mosaics are frequently referred to in discussions of elite life, primarily because they showcase the world of rich African aristocrats. Scholars have suggested varying degrees of realism for the designs on the mosaics. Roger Wilson, for example, argues that generic representations of large houses and grounds were valued forms of iconography, producing desirable statements of wealth and status.24 Katherine Dunbabin, by contrast, views such depictions as actual images of real places, basing her argument on the lack of conventional motifs included in their portrayal.25
Amy Place 182
The true position is likely to lie somewhere between these two interpretations. Clothing, as will be argued below, con-tributed to how viewers interpreted the imagery by framing the iconographic message within dominant cultural ideals.
Both mosaics were excavated from urban settings, yet the mosaic iconography advertises elite otium – a somewhat ambiguous Latin term broadly denoting a sense of leisure or freedom.26 The mosaic pavements transport the viewer into this rural ideal, communicating associated cultural con-notations and core values.27 In this respect, these mosaics were manifestations of a discourse that increasingly defined elite identity. Details of dress in the iconography contributed to this strategy of elite self-representation and thus can be seen as artefacts of an image-clothing rhetoric.
Dominus Julius Mosaic: Reinforcing social distinction through clothing
The desire to visualise aspects of the elite experience on mosaic floors can be seen in North African pavements that directly commemorate the patron’s lifestyle. This ideal was
neatly encapsulated by the well-known Dominus Julius Mosaic which portrays a version of the rural idyll even though found in an urban setting (Fig. 13.1).28 Very little is known about the original archaeological context, although it has been suggested that the mosaic floor once decorated a room with a fountain.29 Dating to around 380–400 CE, this vast mosaic (4.5 × 5.65 m) is composed of three registers of small vignettes organised around an image of a substantial building which dominates the landscape.30 These vignettes anchor the imagery as the representation of the elite habitus by including activities associated with its inhabitants and the estate’s productivity. In the upper left corner, a male servant carries two ducks. There are also peasants gathering olives in a basket. A female servant offers a basket full of olives to the domina who is seated under the shade of the cypress trees (Fig. 13.2). In the same row, another servant brings the domina a lamb while to the right of this scene a shepherd watches his flock. Flanking the estate in the centre is a mounted rider on the left, the dominus. He is accompanied by a servant. On the other side of this impres-sive structure are two hunters who appear to be leaving the
Fig. 13.1. Depiction of the Dominus Julius estate with various scenes of agricultural activities and rural life from Carthage, c. 380–400 CE. Musée du Bardo, Tunis, inv. no. 1. Photo: Sean Leatherbury/Manar al-Athar (CC BY-NC-SA 2.0).
13. Reading dress and identity in the Roman mosaics of Carthage and Tabarka 183
house with their hunting dogs. Imagery in the lower row also reinforces the performance of rural otium. The domina appears again on the left (Fig. 13.3). She leans against a column while her attendant offers her a necklace from a jewellery box. In the far right, a person carries a basket of roses. The other scene in this row depicts the seated dominus who receives a scroll from a figure. Completing this scene is another worker. He carries a basket and grasps a rabbit by the hind legs.
This mosaic, though composed of rather commonplace imagery, has adopted an unconventional format to produce a more personalised and selective approach to mosaic motifs.31 With the use of superimposed levels, the design amalgamates aspects of estate life in one single artistic snapshot. The primary iconographic message is that of the leisurely lifestyle of the rural elite – no doubt due to the wealth of the estate – and this is exaggerated through the repeated depiction of the elite couple alongside their numer-ous workers. The imagery revolves around the dominus and domina and this confirms their superior status.32 The patrons are either approached by the other figures who produce gifts for the domini or, in the case of the middle level, by an attendant who walks besides the dominus who is shown mounted on a horse.
The way the figures in the Dominus Julius Mosaic are dressed would have reinforced social distinction and
reaffirmed the elite position of the patron and his family. In
reaffirmed the elite position of the patron and his family. In